932
Anal. Chem. 1981, 53, 932-934
Dialysis Electrode for Microelectrochemistry Kenneth A. Rublnson Department of Chemistry, University of Cincinnati, Cincinnati, Ohio 4522 1
The general design of an annular electrochemical cell with a porous boundary dates to the work of Tom and Hybbard in 1971 (I). A simpler apparatus was developed by Caja et al. using a porous cation-exchange membrane (2). A number of limitations were noted by the authors. Described below is a similar, simple apparatus (Figure 1) which overcomes many of the drawbacks of the earlier designs. It comprises a wire working electrode surrounded by a tubular cellulose acetate dialysis membrane as a boundary for the electroactive species (but not for the background electrolyte). EXPERIMENTAL SECTION Figure 1is a pictorial diagram of the working electrode. Details of ita construction are described in the figure caption. The syringe used is airtight with a volume of 10 pL (Precision Sampling Corp., Baton Rouge, LA, Model C160). The apparatus was used as a working electrode by immersing it in a 10-mL beaker with a Pt-wire auxiliary electrode, calomel reference electrode, and capillary tube for argon bubbling. AU voltammograms were obtained while argon was passing through the solution unless noted otherwise. This aeration stirs the external solution permitting faster equilibration across the membrane as well as keeping the system oxygen free. The beaker could be left open to the atmosphere for easy access. All the deoxygenated solutions used were made so by argon purging before use and then stored in closed bottles. The sample to be investigated is held in the annulus between the electrode and the surrounding membrane. The surrounding solution reservoir contents could be changed easily. The ionic strength and pH are assumed to be the same in the reservoir and in the annulus. To counteract osmotic effects tending to dilute the electroactive species, the solution in the external reservoir was made with an equal concentration of Carbowax 1000 (Union Carbide) which is a poly(ethy1ene glycol) fraction of average molecular weight 1000. The poly(ethy1ene glycol) solution decomposes within a few days, so a new stock solution (100 mM; 1 g/10 mL) was made each day with doubly distilled water. The pH of the solution was measured while the cyclic voltammogram was being run. This was done by using a Fisher combination pH electrode with an Orion Model 601A pH meter. The equilibrium hydrogen ion concentration was assumed to be the same on both sides of the membrane. The pH buffers were made from stock 0.5 M preelectrolyzed AR sodium phosphate and phosphoric acid solutions. Preelectrolysis was carried out at a mercury pool for at least 24 h. The gold wire (99.9%) and platinum wire (99.9%),both 0.5 mm diameter, were purchased from Alfa. The 2,6-dichlorophenolindophenol sodium salt (Purim),abbreviated throughout as DCIP, was purchased from Fluka. A stock solution was stored at pH 7 in phosphate buffer. Without buffer, the compound decomposed within days to a red material that precipitates at neutral pH. The voltammograms were obtained by using a Bioanalytical Instruments, Inc., Model CV 1A potentiostat. Unless stated otherwise, a scan speed of 20 mV/s was utilized. The potentials were measured from the output with a Fluke 8000A multimeter. Peak potentials were found from interpolation on the graph paper of the x-y recorder (Hewlett-Packard Model 7004B) and are accurate to 5 mV. Differential pulse voltammetry was done by using a Princeton Applied Research Model 174A polarograph. The conditions used were: scan rate 10 mV/s, modulation amplitude 10 mV, pulse frequency 2 Hz with no output filtering. No precautions were taken to deoxygenate the solution. For the 10 pM solution, no external polymer was used to balance osmotic differences across the membrane. The gold electrode wire was cleaned with concentrated nitric acid prior to the experiment. The membrane was mounted afterward. The membranes used on the working electrode assembly were composed of cellulose acetate (Eastman Tenite 008A) which was
formed into tubes by FRL, Inc., Dedham, MA. As received, the tubing has a diameter of 690 pM inside with a wall thickness of 30-34 pm. After hydrolysis, the size can change slightly with pH and hydration changes (3). The originally impermeable tubing was made porous on the molecular level by hydrolysis at ambient temperature (18-20 O C ) with 0.2 N KOH simultaneously inside and outside. For the compounds investigated here (mol w t = 300), the membranes were hydrolyzed for 6 h. One end of each of the 3-5 cm long sections was left unhydrolyzed to make mounting easier. After the tubes were hydrolyzed, they were washed with distilled water. They could be stored without apparent changes in their properties for at least a month in water.
RESULTS AND DISCUSSION The novel properties of the apparatus depend mostly on the membrane properties. Different hydrolysis times cause the membrane material to exhibit different selectivities of molecular size. For example, after hydrolysis for 2 days, the tubing is freely permeable to organic molecules up to molecular weight 1000. However, after a 4-h hydrolysis, the tubing is permeable to protons as seen by immersing a tube in acid with a pH indicator inside. But after the 4-h treatment, the membrane still had too high a resistance to use in running voltammograms. The tubes hydrolyzed for 6 h held larger electroactive species (e.g., tetraphenylporphyrins) with molecular weights above 600 indefinitely. Of course the selectivity of the membrane is based on the effective molecular size, and so small inorganic coordination complexes with higher molecular weights (e.g., containing ruthenium) freely passed through tubes treated for 6 h. The tubing used here, which was hydrolyzed for 6 h, was freely permeable to small background electrolyte ions and retained organic molecules of molecular weight 300 with a half-time (at the electrode) of 2-4 min. This lower molecular weight range is more difficult to work in, and so the examples shown here are in that range. In Figure 2 are shown the results of an experiment where, with one sample, the compounds Eo'values were determined at a number of pH values. The total time elapsed for the set of voltammograms to be made was less than 20 min. In the experiment, the DCIP solution was 10 mM and the sample volume was about 2 pL. Thus the load consisted of 20 nmol of the compound. This relatively high concentration was used to allow more pH changes. As shown below for chlorpromazine, much lower concentrations are usable. Depending on the size of the pH jump, the sample reached its new equilibrium 0.5-2 min after the external solution was changed to a new pH. This was determined by the stabilization of the voltammogram over time. Because of the membrane-electrode geometry, a number of novel benefits arise as seen in this experiment. First, the study was carried out in a 10-mL beaker open to the air with argon gently bubbled through the solution outside the membrane. This procedure keeps the whole system deoxygenated. As a result, the sample need not be deoxygenated before loading it into the cell. The hundred-micron-thick annulus containing the sample solution equilibrates in seconds with the argon-purged reservoir. Second, since the membrane prevents convection near the electrode, no delay need be made to allow the cessation of convective flow. As a result, the voltammograms of the samples can be run almost as fast as they can be loaded.
0003-2700/81/0353-0932$01.25/00 1981 American Chemical Society
ANALYTICAL CHEMISTRY, VOL.
1
1
0
,
I
( I I (
Flgure 1. Diagrammatic cross section of tho membrane-bounded working electrode. This is Immersed to the top of the membrane In a 10-mL beaker with the auxiliary and reference electrodes: (A) 0.5-mm Au or pt wire; (B) a section of melting-point capillary used as a connector; (C) cellulose acetate membrane; (D) polyethylene capillary tubing leading to the, mlcrosyringe used for loading the sample. When a sample Is loaded, the lower tlp of the membrane tube is left filled with air, which acts aa an insulator. The seals (stippled areas) are dentists' "brown sticky wax", but epoxy can be used for more durability.
